MiniMax: MiniMax M3: AI model
MiniMax M3 is a native multimodal model with a 1M-token context and sparse attention for cheap long-context work. Accepting text, image and video, reachable from inside Iran through Hormouz.
- Context window:
- 1,048,576 tokens
- Input:
- $0.3/1M tokens
- Output:
- $1.2/1M tokens
- Type:
- text, image, video
- Vision
- Tools
- Streaming
- Long context
Thinking levels:
MiniMax M3 is a native multimodal foundation model that accepts text, image and video input and produces text output, with a 1M-token context window. It is aimed at long-horizon agentic work, coding and tool use.
It is built on MiniMax Sparse Attention, which replaces full attention with block-based key/value selection to cut per-token compute at long context. In practice this means much cheaper and faster processing at 1M tokens while keeping quality across most tasks, and it was tuned for multi-turn, production-style collaboration.
Pricing is usage-based: $0.3 per million input tokens and $1.2 per million output tokens. Hormouz provides stable access to MiniMax M3 from inside Iran, with no foreign credit card or circumvention tooling required.
MiniMax-M3 is a multimodal foundation model from MiniMax. It supports text, image, and video inputs with text output, a 1M-token context window, and is suited for long-horizon agentic work, coding,...
Quickstart
curl https://api.hormouz.net/v1/chat/completions \
-H "Authorization: Bearer ng-YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "minimax/minimax-m3",
"messages": [{"role": "user", "content": "Hello!"}]
}'Specs
| Vendor | MiniMax |
|---|---|
| Type | Text, Image, Video |
| Context window | 1,048,576 tokens |
| Pricing | $0.30/M in · $1.20/M out |
| API | Chat Completions |
| Capabilities | Streaming · Vision · Tools |
Benchmarks
Reasoning
Coding
Agentic
Source: Artificial Analysis
Design Arena
How Design Arena works ↗Metrics sourced from Design Arena
Code example
curl https://api.hormouz.net/v1/chat/completions \
-H "Authorization: Bearer ng-YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "minimax/minimax-m3",
"messages": [{"role": "user", "content": "Hello!"}]
}'from openai import OpenAI
client = OpenAI(api_key="ng-YOUR_API_KEY", base_url="https://api.hormouz.net/v1")
response = client.chat.completions.create(
model="minimax/minimax-m3",
messages=[{"role": "user", "content": "Hello!"}],
)
print(response.choices[0].message.content)import OpenAI from "openai";
const client = new OpenAI({ apiKey: "ng-YOUR_API_KEY", baseURL: "https://api.hormouz.net/v1" });
const response = await client.chat.completions.create({
model: "minimax/minimax-m3",
messages: [{ role: "user", content: "Hello!" }],
});
console.log(response.choices[0].message.content);Buy the MiniMax: MiniMax M3 API, paid in rial
- Buy credit with an Iranian bank card; no foreign card or account needed.
- One Hormouz key reaches MiniMax: MiniMax M3 and 600+ other models, and you pay only for what you use.
- No VPN and no risk of a banned account; just change the base URL to api.hormouz.net/v1.
Use via Hormouz
Connect to this model and 600+ others with one API key, then see usage down to each request.
Request company accessFAQ
- How much does MiniMax M3 cost?
- Billing is usage-based: $0.3 per million input tokens and $1.2 per million output tokens. There is no fixed monthly fee. You pay only for what you use, at the rates shown in the Hormouz dashboard.
- How do I access MiniMax M3 from inside Iran?
- Create an account on Hormouz and generate an API key. Hormouz provides a stable path to MiniMax M3 from inside Iran, with no foreign credit card or circumvention tooling required.
- Is MiniMax M3 compatible with OpenAI code?
- Yes. Hormouz exposes an OpenAI-compatible API. Just change the base URL to Hormouz and swap in your Hormouz key. Your existing OpenAI SDKs and code run without a rewrite.
- Does MiniMax M3 support images and video?
- Yes. M3 is a native multimodal model that accepts text, image and video input with text output. It uses sparse attention to make its 1M-token context cheaper and faster, and is tuned for long-horizon agentic and coding work.
Updated 2026-10-13